Heat: high-efficiency simulation for thermal ablation therapy.

Area of Science:

• Medical Physics
• Computational Biology
• Image-Guided Therapy

Background:

• Percutaneous thermal ablation is a widely used minimally invasive procedure.
• Preoperative planning for thermal ablation, especially with multiple needles, is complex and time-consuming.
• Current planning methods involve trade-offs between speed and accuracy in thermal propagation simulation.

Purpose of the Study:

• To develop and validate a multi-resolution approach for accelerating thermal propagation simulation in percutaneous thermal ablation.
• To enable interactive adjustment of ablation parameters and real-time visualization of predicted ablation volumes.
• To improve the efficiency of preoperative planning for procedures involving multiple ablation needles.

Main Methods:

• A GPU-accelerated multi-resolution simulation combining high-resolution (finite difference/lattice Boltzmann) and low-resolution estimations.
• Comparison of finite difference and lattice Boltzmann methods for GPU acceleration.
• Parameter study to optimize the balance between speed and accuracy for different resolution frames.
• Validation in multi-needle scenarios for interactive planning.

Main Results:

• The multi-resolution method significantly reduces computation time compared to reference simulations.
• High-resolution frames achieve up to 5.8 fps, while low-resolution frames reach 32 fps with <20% accuracy loss.
• The approach maintains good accuracy for predicting ablation volumes.

Conclusions:

• The multi-resolution approach facilitates smooth, interactive planning for percutaneous radiofrequency treatments with multiple needles.
• Instant visualization of predicted ablation volumes enhances user experience and planning efficiency.
• This method holds potential for automated treatment planning, reducing iterative adjustment times.